[Present and future of trial design]

Nowadays, clinical researchers can choose from a large variety of designs. These can be classified as observational or experimental, descriptive or comparative, longitudinal or cross-sectional, prospective or retrospective, "blind" or "unblind", prolective or retrolective. The randomized double-blind trial is still the most controlled design and, therefore, the least susceptible to bias. However, ethical, economical and generalizability reasons limit their use. Current tendencies are towards the use of poorly controlled studies that can provide weak evidence and could, ultimately, do more harm than good to our patients and our economy. There is, therefore, the need for designs that could I) provide high quality evidence, II) substitute an experiment when it cannot be done, and III) overcome the limitations of randomized clinical trials.